torch.nn.quantized.FXFloatFunctional#

class torch.nn.quantized.FXFloatFunctional[源代码]#

module to replace FloatFunctional module before FX graph mode quantization, since activation_post_process will be inserted in top level module directly

Valid operation names:

  • add

  • cat

  • mul

  • add_relu

  • add_scalar

  • mul_scalar

上一页

torch.nn.quantized.FloatFunctional

下一页

torch.nn.quantized.QFunctional